[Ozone] Initialize SurfaceFactoryOzone only in the EGL GLES2 case

[Ozone] Initialize SurfaceFactoryOzone only in the EGL GLES2 case

aura_unittests explicitly calls GLSurface::InitializeOneOffForTests in main().
The Ozone platform isn't initialized this early causing a CHECK failure when
trying to get SurfaceFactoryOzone.

At this point surface creation in unittests is handled by OSMesa, thus we don't
need SurfaceFactoryOzone initialized.

TBR=piman@chromium.org
NOTRY=true

TESTS=Ran aura_unittests manually using an Ozone build

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=260239

[dnicoara, 2014-03-28 19:22:06]

The CQ bit was checked by dnicoara@chromium.org

[commit-bot: I haz the power, 2014-03-28 19:25:21]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/dnicoara@chromium.org/217003007/1

[commit-bot: I haz the power, 2014-03-28 19:29:50]

Change committed as 260239

[piman, 2014-03-28 19:57:38]

https://codereview.chromium.org/217003007/diff/1/ui/gl/gl_surface_ozone.cc
File ui/gl/gl_surface_ozone.cc (right):

https://codereview.chromium.org/217003007/diff/1/ui/gl/gl_surface_ozone.cc#ne...
ui/gl/gl_surface_ozone.cc:33: break;
actually, should we return false here?

[dnicoara, 2014-03-28 20:30:36]

https://codereview.chromium.org/217003007/diff/1/ui/gl/gl_surface_ozone.cc
File ui/gl/gl_surface_ozone.cc (right):

https://codereview.chromium.org/217003007/diff/1/ui/gl/gl_surface_ozone.cc#ne...
ui/gl/gl_surface_ozone.cc:33: break;
On 2014/03/28 19:57:38, piman wrote:
> actually, should we return false here?

As is no since OSMesa uses this path and we need to return true for that. Maybe
we should add another case statement for OSMesa then we can return false for
everything else.

[piman, 2014-03-28 20:48:56]

On 2014/03/28 20:30:36, dnicoara wrote:
> https://codereview.chromium.org/217003007/diff/1/ui/gl/gl_surface_ozone.cc
> File ui/gl/gl_surface_ozone.cc (right):
> 
>
https://codereview.chromium.org/217003007/diff/1/ui/gl/gl_surface_ozone.cc#ne...
> ui/gl/gl_surface_ozone.cc:33: break;
> On 2014/03/28 19:57:38, piman wrote:
> > actually, should we return false here?
> 
> As is no since OSMesa uses this path and we need to return true for that.
Maybe
> we should add another case statement for OSMesa then we can return false for
> everything else.

Yeah, I think that's what we want. Explicitly whitelist the ones we support, and
fail the rest.